BigQuery Data Connectors for Ecommerce Businesses

It’s a Monday morning. Shopify reports $48k in sales last week. Meta has $31k in purchases, Google Ads has $22k and Klaviyo has another $15k. Your channels are showing you more revenue than your store has actually made.

Every ecommerce team will have come across this. Each platform has its own way of measuring revenue and none of them have the full picture. The answer is to bring your store data and your marketing data together. That then allows you to compare the claims made by each channel to the actual orders and apply your own rules to work out what your revenue is.

BigQuery is a fairly common tool for this. The real challenge is working out how to get your data into it. This guide looks at three different connectors, each with a different methodology.

Key Takeaways

• Revenue reported by each platform is very rarely going to match your store’s total as each platform has its own attribution rules. A data warehouse helps in that it allows you to bring the orders together with the ad spend and create one single source of truth.
• Your connector dictates where the work is actually going to happen. Some tools will clean and structure the data prior to loading it; others will simply load the data as it is, and then you’re responsible for the modeling.
• Where your data is stored is just as important as how it gets there. If you’re using your own BigQuery project, you have access to all the tools and AI assistants. A vendor’s warehouse is obviously quicker to set up but once you’ve got it in place it’s much harder to get out of.

Five Questions to Ask Before You Pick a Connector

1. Does it cover your store in any detail at all? Orders by themselves aren’t sufficient. You need line items, refunds, customers and products if you’re going to work out margin and repeat purchase rate.
2. Does it bring in marketing data as well? A connector that only deals with the store answers half your questions. Ad data, email data and analytics data all need to be sitting alongside your orders.
3. Who cleans the data? Shopify orders are nested and refunds come in as separate records. Someone has to flatten and blend them (either the connector or your SQL).
4. Whose warehouse is it? Some of the tools will load the data into your own BigQuery project. Others will keep it in theirs. That has a massive impact on ownership, cost and what you can then connect to later on.
5. Can you ask it questions? Increasingly more teams are using AI assistants to analyze store data. Check that the connector can make warehouse data available to tools such as Claude or ChatGPT.

Three Ways to Get Ecommerce Data into BigQuery

Coupler.io and Skyvia both move ecommerce data into BigQuery; the way they do it is quite different. Polar Analytics is another option for brands who want ready made analytics and don’t want to go to the trouble of setting up their own BigQuery pipeline.

1. Coupler.io: Prepare the Data, Then Ask AI About It

Best for: Ecommerce brands that want BigQuery and AI analytics without a data engineer.

Coupler.io is a no-code data integration and AI analytics platform with 400+ sources. For an online store, it would cover Shopify, WooCommerce, Stripe, Klaviyo, GA4, and the major ad platforms. Data loads into your own BigQuery project on a schedule.

The key difference is in the data cleaning. You can filter fields, rename columns and join orders with ad spend prior to loading into BigQuery. BigQuery then gets tables that already answer the ‘what did each channel actually bring in?’ question for Monday morning.

Then Coupler.io takes the next step. Through its MCP server, it connects BigQuery data to AI assistants such as Claude, ChatGPT, and Gemini. You can also connect Shopify to AI directly for quick store questions.

The AI takes your question and turns it into an SQL query, which is then run against your actual data in the warehouse. The totals are derived from the warehouse, not calculated by the AI. Pricing is based on the number of accounts you connect. The free plan includes data loading and the AI features kick in from the Professional plan onwards.

Example: A founder asks Claude, “Which product categories had the highest refund rate among customers who came from paid social?”

Trade-off: Coupler.io is built for business apps and marketing data, not for database backups. If you’re looking to backup your production databases then you’ll need a different tool.

2. Skyvia: Replicate Your Store Data, Then Model It

Best for: Teams with SQL skills who want a complete raw copy of their store data.

Skyvia is a no-code cloud data integration platform with 180+ connectors (Shopify, BigCommerce, Recharge, Yotpo etc). Its replication tool will replicate store objects into BigQuery and create the tables for you. It supports incremental updates so your copy stays up to date and new source fields will appear in BigQuery as they are added.

It’s effectively a mirror of your store. You select the objects and fields you want to copy, you can add filters if you wish and the data appears in BigQuery in much the same way as it does in your store. For more complex changes you can use Skyvia’s Import and Data Flow tools. There is a free plan and the paid plans are based on features and volume of data.

Example: A subscription brand is replicating all of their Shopify and Recharge objects into BigQuery and then building their churn models in SQL.

Trade-off: the mirror gives you all the data but no explanations. You’ll need to blend orders with ad spend and calculate your ecommerce metrics in SQL after the load, so marketing reports are going to require a fair bit of modeling work.

3. Polar Analytics: Skip the Warehouse Work Entirely

Best for: Shopify brands who want fully finished dashboards and attribution in days.

Polar Analytics is a completely different proposition. It’s an ecommerce analytics platform (primarily for Shopify) with over 45 sources (Shopify, Amazon, Meta, Google, TikTok, Klaviyo etc). You get a range of pre-built metrics straight out of the box (contribution margin, LTV by cohort, etc.) and multi-touch attribution via their own server-side pixel.

There’s a caveat for BigQuery users in that Polar doesn’t load data into your BigQuery project. You each get your own Snowflake warehouse, and SQL access is plan-dependent. Polar can also pull in your existing BigQuery data as a source and has an MCP connection for Claude and ChatGPT.

Example: A Shopify brand with no data team wants their profit and attribution dashboards ready before the next board meeting.

Trade-off: if you’re already using BigQuery as your warehouse, then Polar will create a second warehouse rather than use yours. Pricing is based on your online GMV and starts at around $400/month, which can be a bit of a cost for smaller stores.

Which Approach Fits Your Store?

Start by asking yourself do you want your own warehouse? If you do, then pick a tool that loads data into BigQuery. If you’re primarily after dashboards and attribution then Polar is by far the quickest option.

If you go with BigQuery, then the next thing to think about is who’s going to do the modeling. If you have someone on your team who knows SQL then the Skyvia raw copy gives them complete control. If you don’t, then Coupler.io does the modeling for you and adds in the AI analytics.

For the majority of growing ecommerce brands, the last option is the one that makes the most sense. You retain ownership of your data, your tables are ready for reporting and you can just ask questions in plain English rather than having to wait for a new dashboard.

FAQ

Why not just use Shopify’s built in reports? The Shopify reports are good at covering store activity but don’t include ad spend, email or any of the other channels. In BigQuery you can join the orders to the marketing data and keep the history as long as you want.

How much does BigQuery cost for an online store? The free tier is 10GB of storage and 1TB of queries per month. A small store with fairly efficient queries can easily stay within that. Costs go up with both storage and query volume so you’ll need to watch that as your reporting increases.

Do I need a data engineer to use these connectors? Not to set them up (all three are no-code). SQL is the key skill when it comes to raw replication where the metrics are calculated post load.

Can AI analyse my store data in BigQuery? Yes it can. Coupler.io will connect your BigQuery tables to assistants such as Claude and ChatGPT. You can then check the AI’s answers against the source totals (particularly useful for revenue and refunds).